Intergroup biases
1
Intergroup Biases in Conflict:
Reexamination with Arab Pre-adolescents and Adolescents
Yona Teichman
Daniel Bar-Tal
Yasmina Abdolrazeq
Tel-Aviv University
We thank Elana Fein and Elana Kleiner-Lowell for their valuable comments.
Correspondence concerning this article should be addressed to Yona Teichman Department of
Psychology, Tel-Aviv University, Tel-Aviv 69978, Israel. E-mail: yonat@post.tau.ac.il
Intergroup biases
2
Abstract
This study examined the proposition derived from the integrative developmental contextual
theory (IDCT) (Bar-Tal & Teichman, 2005; Teichman & Bar-Tal, in press) that contextual
circumstances determining collective self esteem (i.e. conflict and social status), and
developmental stage in which identity development is a central issue (pre- and early
adolescence) would influence intergroup biases. This proposition was previously examined
with Jewish Israeli participants and here is reexamined in three samples, aged 8 – 17 including
Israeli Arabs and two groups from the Palestinian Authority. The three samples were based in
different intergroup contexts in which their group status was relatively high, medium or low.
The findings demonstrate developmental differences in the relationship between contextually
defined collective self-esteem and intergroup attitudes. Results confirmed the hypothesis that
in conflict, irrespective of age participants' intergroup biases would reflect group status.
However, the hypothesis that the lowest status group would express only more negativity
toward the out-group was not confirmed. The developmental hypothesis that pre- and early
adolescents from the high status group would express most varied and extreme biases was
also confirmed. With some exceptions in the middle status group, 14 – 17-year-olds displayed
moderation in biases.
Intergroup biases
3
Intergroup Biases in Conflict:
Reexamination with Arab Pre-adolescents and Adolescents
The focus in this article is on intergroup biases expressed by pre-adolescents and adolescents
who grow up in an intractable conflict. The conflict, which is the socio-psychological context
of this study is known as the "Israeli – Arab conflict". Most of our previous research on
stereotypes and prejudice in conflict was conducted with Israeli Jewish children and
adolescents revealing polarized beliefs and attitudes toward in- and out-group and often
peaking biases during pre- and early adolescence (Bar-Tal & Teichman, 2005; Teichman,
2001). Given the possibility to access Arab communities we have the unique opportunity to
examine some of our theoretical propositions regarding the development of stereotypes and
prejudice with participants experiencing the same conflict from the other side. For these
participants the Israeli Jews represent the out-group and enemy.
In the last two decades, the examination of the developmental trajectory of intergroup
stereotypes and prejudice has drawn on socio-cognitive theory (S-CT) (Aboud 1988), which
contends that cognitive development achieved at the age of 7 – 9 leads to the reduction of
prejudice.. Following this line of thinking further cognitive development occurring at the age
of 10 and later should moderate the level of prejudice. However, studies conducted in nonviolent societies and in conflict that included pre- and early adolescents reported a renewed
elevation in prejudice during this age (Black-Gutman & Hickson, 1996; Bar-Tal & Teichman,
2005; Rutland, 1999; Teichman, 2001; Vaughan, 1987). Since S-CT focuses on the
development of prejudice in preschoolers and school-age children, and so does most of the
research it generated, this theory does not offer a theoretical explanation for the findings
reported for pre- and early adolescents.
On the other hand, the integrative developmental contextual theory (IDCT) we have
proposed (Bar-Tal & Teichman, 2005; Teichman & Bar-Tal, in press) suggests that
Intergroup biases
4
stereotypes and prejudice develop in context and at any given time are influenced by affect,
cognitive development, and identity development each gaining salience in a different
developmental stage. IDCT covers a developmental span from 2 – 17 and presents a
developmental view of stereotypes and prejudice that includes different stages of adolescence.
In adolescence, 12-years and later (the stage of formal operational thought in Piaget’s
terms) abstract and hypothetical thinking is acquired. The advance in cognitive development
and acquisition of broader and more flexible perspective taking opens adolescents to relate to
moral, social and political issues (Kohlberg, 1969; Piaget, 1952; Selman, 1980). Further
development, in later stages of adolescence (14 and later) is reflected in valuing justice,
dignity, equality and human rights. This approach to social-cognitive development suggests
that adolescents' intergroup biases should decrease (for an alternative view, see Killen, LeeKim, McGlothlin & Stangor, 2002; Killen, Margie & Sinno, 2006). However, apparently,
social-cognitive development constitutes a potential, the realization of which depends on other
factors that acquire critical influence in this developmental stage. IDCT suggests that these
factors are identity development and the context in which development occurs.
From pre- and early adolescence (ages 10 - 13 and later), identity formation and
consolidation become the main developmental task (Erikson, 1968; Marcia, 1980; 1989). The
development of the integrated self-identity includes different aspects of identity from which
personal-identity, collective-identity, and the self-esteem related to each are of interest here 1.
The process of identity formation begins with self-redefinition confronting adolescents with
questions such as who they are and who they want to be. Marcia (1998) points out that the
"changes attended from puberty and in subsequent adolescence have two effects: they
disequilibrate the fairly structured life of the latency-aged child; and they require an
1
Other aspects of self-identity i.e. gender-identity, religious identity, family identity
etc. are also parts of the integrated identity, but are not considered in this article.
Intergroup biases
5
accommodation in the form of a new integration of physique, sexuality, ideology, abilities,
personal needs, and social demands" (p.36). In this developmental stage the status of the
groups to which one belongs and the drive for self-enhancement identified by social identity
development theory (SIDT: Nesdale, 1999; 2000, Nesdale, Durkin, Maass & Griffiths, 2005;
Nesdale & Flesser, 2001) are highly relevant. In-group and out-group, acceptance or rejection
serves to enhance personal and collective worth.
IDCT contends that insecurity related to the process of disequilibration of identity during
pre- and early adolescence results in a high need for self reassurance and this need intensifies
self-enhancement motivation which in turn incites intergroup biases. At the age of 14 - 15,
through late adolescence (age 16 - 17) when gradually a more secure identity is achieved
moderations in biases may be expected to appear. It follows that in the different phases of
adolescence the progression in cognitive development alone can not account for social biases.
Importantly, here we do not take a position in the controversy over whether personal (Wills,
1981) or collective self-esteem (CSE: Crocker, Baline & Luhtanen, 1993; Turner, 1999)
determines intergroup biases, rather, we propose to view social biases in adolescence as
related to the insecurity they experience due to the developmental tasks they face with the
integration of all aspects of their identity. We relate to both personal and social identity, but
because of the context of our research the focus is on social identity and the CSE related to it.
Findings that self-esteem drops in adolescence (Robin, Trzesniewski, Tracy, Gosling, &
Potter (2002) and that pre- and early adolescents manifested an increase in in-group
preference and out-group rejection (Black-Gutman & Hickson, 1996; Bar-Tal & Teichman,
2005; Rutland, 1999; Teichman, 2001; Vaughan, 1987) may be seen as an indirect support for
the influence of the need for self-enhancement. So are findings reported by Nesdale, Maass,
Durkin, Griffth (2005) and Nesdale and Brown (2004) showing that participants aged 9 and
Intergroup biases
6
12 were more sensitive to a negative representative of their in-group than were younger
children.
A more direct examination of the association between assessed personal self-esteem, and
social biases in adolescents (Sasson, 2003, reviewed by Bar-Tal & Teichman, 2005) indicates
that low-self-esteem participants displayed unstable in-group and out-group response patterns
and particularly those aged 10 -12, expressed higher in-group positivity and out-group
negativity. Sasson's (2003), findings for the 10 - 12-year-olds support the theoretical corollary
suggested by Abrams and Hogg (1988) that low or threatened personal self-esteem leads to
social biases and the conceptualization regarding the relationship between developmental
processes and intergroup responses in adolescence proposed by IDCT. These however, are
initial results that need to be further examined.
As noted, stereotypes and prejudice are molded not only by stage-related developmental
experiences, but also by the specific social contexts in which they develop. It may be
suggested that since identity development is also associated with the context, the influence of
the context in adolescence is even more decisive. It is also plausible to assume that the
specific social context of conflict in which the reported study was conducted brings to the
forefront intergroup relationships and collective identity. Turner, (1999) who used the concept
of "share identity" proposed that in real life situations such as conflict where common fate,
shared threat, proximity, similarity, cooperative interactions, or positive interdependence are
salient components, awareness of shared social identity becomes obvious and widely
recognized. In this type of context it is more likely that collective self-esteem will be most
relevant and have a decisive influence on identity development and consequently on the
intergroup biases held by adolescents. As suggested by Crocker and Luhtanen, (1990)
intergroup biases may represent an attempt to enhance the in-group; they may serve the
collective rather than personal self-esteem. Thus, in a conflict situation, where group
Intergroup biases
7
distinctiveness, commitment, and identification increase the enhancement may be associated
more with collective than with personal self-esteem.
In the present study, our conceptualization of CSE should not be confused with the
specific operationalization of CSE at the individual level (see Crocker, Blaine, and Luhtanen
(1993). CSE in our research is conceived as a group characteristic that involves the social
status, history and intergroup relationships in which particular groups areembeded. Thus, it
embraces social standing of the group as a whole in a larger sense rather being a measure of
individual differences. Viewing CSE as inferred by individuals from the status of their group
in the society overcomes a problem of previous research with CSE identified by Long and
Spears (1998). According to them "one problem with the previous research is that it has
tended to view CSE as a stable, cross group personality trait variable rather than as contextdependent group-specific state variable" and further, "It seems unlikely that the esteem in
which a social identity is held can be reduced to a stable individual difference variable….CSE
is contingent upon the positivity of a social identity, which in turn varies according to aspects
of the context such as the relevant out-group or the dimension of comparison." (p. 916).
In the present study, we assume that individuals are aware of the social evaluation of their
groups and integrate it into their identity. Thus, in this study, CSE was neither assessed by an
individual self-report measure nor manipulated, rather it was determined by two parameters
rooted in the cultural context of the examined groups and accepted in the general society and
in their respective groups. The parameters we applied for differentiating among the examined
groups were exposure to the conflict, and social-economic status (SES). The group with high
exposure to the conflict and high SES was defined as high in CSE (HCSE). The group with
low exposure to the conflict and low SES was defined as low in CSE (LCSE), and a group
with high exposure to the conflict and very low SES was defined as a group with moderate
collective self-esteem (MCSE).
Intergroup biases
8
Introducing the concept of CSE at the individual level Crocker, Blaine, and Luhtanen
(1993) suggested that when threatened, in real meaningful situations individuals with high
collective self esteem (HCSE) will enhance their self-worth expressing in-group preference
and out-group prejudice. On the other hand, due to the need to preserve self-consistency,
individuals with low collective self-esteem (LCSE) will protect their self-worth by derogating
the out-group i.e. by prejudice. Integrating the propositions regarding the influence of CSE on
biases with IDCT's developmental propositions regarding adolescence suggests the following
hypotheses: 1. Irrespective of age HCSE group will manifest highest in-group positivity and
lowest in-group negativity than the two other groups. Compared with the HCSE group, the
LCSE group will manifest same or higher out-group negativity and same or lower out-group
positivity. The MCSE will hold an in-between position. 2. Developmentally: Pre and early
adolescents will express peaks of in- and out-group positivity/negativity. The predicted age
pattern will be most clearly manifested by pre- and early adolescents in the HCSE group.
LCSE pre-and early adolescents will express mainly low peaks in out-group positivity and
high peaks in out-group negativity. However, because of their unclear identity status pre- and
early adolescents in the MCSE group are expected to resemble the LCSE group.
To examine these hypotheses we defined CSE at the group level as a predictor of
intergroup biases and assessed positive and negative trait attributions to in- and out-group,
positive and negative feelings toward the out-group and readiness for social contact withoutgroup members. Positive/negative attributions to the in-group, and to the out-group were
analyzed separately, so were positive/negative feelings toward the out-group.
Method
Participants
The participants in this study were Arab children and adolescents representing different social
contexts. The first group was comprised of Israeli Arab citizens from the city of Taiibe. The
Intergroup biases
9
two other groups were comprised of Palestinians from the Palestinian Authority, one from the
city of Hebron and one from Al-Fawara refuge camp. The Israeli participants are
characterized as a LCSE group. They belong to a discriminated stably low status minority
group within the State of Israel, with impermeable borders between them and the Israeli
Jewish majority. They grow up with a dual identity. On the one hand they enjoy the benefits
of the Israeli citizenship, but at the same time share an ethnic and ideological bond with the
Palestinians in the Palestinian Authority, many of them prefer to be identified as Israeli
Palestinians. Socio economically, they represent low and middle class. They do not serve in
the Israeli army, and only very rarely encounter conflict-related violence.
The two other groups have a clear Palestinian identity, strong commitment to the Palestinian
cause. The group from Hebron classified as HCSE group lives in a divided city in which
major parts belong to the Palestinians and some quarters to the Jews. In Hebron, conflict and
confrontations are daily occurrences creating a very high probability to witness or experience
violence. The participants from Hebron represent a proud middle-class group with long
history of rivalry with the Jews that may be traced to times before the establishment of the
State of Israel. The second group from the Al-Fawara refugee camp is the MCSE group. They
are under Israeli occupation which implies high exposure to conflict-related violence.
However, they belong to a disadvantaged low-class enrooted from their homeland, not
accepted as equals by their current environment and disrespected by their neighbors.
All together there were 852 participants aged 8 - 17. The LCSE group included 337
participants, 72 were 8 – 9, 74 were 10 - 11, 66 were 12 - 13, 46 were 14 - 15, and 79 were 16
- 17. The HCSE group included 195 participants, 56 were 8 - 9, 42 were 10 - 11, 48 were 12 13, 30 were 14 - 15, and 20 were 16 – 17. The MCSE group included 313 participants 62
were 8 - 9, 74 were 10 - 11, 74 were 12 - 13, 54 were 14 - 15, and 55 were 16 - 17. All three
Intergroup biases 10
samples were composed from a similar number of males and females. In a preliminary
analysis gender did not reveal any systematic differences.
Measures and Procedure
Stereotyping was assessed by trait attributions to Jews and Arabs. The specific traits were
determined by a pilot study, in which 150 children and adolescents were asked to list the most
salient traits of Arabs and Jews. Traits that were noted by at least 80% of the respondents for
each social category were included in the rating list. The final list included 18 traits, nine from
those identified for Arabs and nine for Jews. In order to include equal number of positive and
negative traits, occasionally opposites were used. The specific traits were: clean, good, thief,
smart, terrorist, lazy, beautiful, fresh, stinking, strong, educated, liar, cruel, loyal, hospitable,
coward, sociable, and violent. These traits were separated into positive and negative ones.
Rating was performed on a 5 point scale indicating the incidence of each trait for each target
group: (1) none of them –0%, (2) few-25%, (3) half-50%, (4) many-75%, and (5) all of them –
100%, as such it also serves as a homogeneity measure. The measure yielded the following
Cronbach alphas: For the positive traits attributed to Jews .79 and to Arabs .87, for the
negative traits attributed to Jews .79 and to Arabs .87.
Attitudes were expressed in feelings toward the Jews and in indicating readiness for social
contacts with them. The measure of positive feelings included: liking, pity, closeness, and
empathy, and the measure of negative feelings included: hate, fear, disgust, anger. The
responders indicated whether they experienced the particular feeling (3), experienced it
moderately, (2) or did not experience it at all (1). The mean range for the positive and
negative feelings was 1 – 3. Cronbach alpha for the positive feelings was .70 and for the
negative ones .80.
Readiness for social contact with “a Jew of your age and gender who speaks Arabic” was
indicated for three possible encounters: to meet with him/her, to host him/her in your home,
Intergroup biases 11
and to befriend him/her. The responses for each possibility were yes/no and provided a scale
from 0 (not ready for any contact) to 3, (ready for the three types of contact).
Participants answered the questionnaires anonymously at school in their classes. The time
the data was collected was relatively calm, with no outbursts of violence between Jews and
Arabs.
Results
Comparisons of the trait attributions by the 5 age-groups from the three different CSE groups
were performed with repeated measures three way ANOVAs. Different analyses were
performed for measures indicating positivity and negativity. The independent variables were
target (Jew/Arab), responder identity (3 CSE groups), and age (5 age groups). Feelings and
social distance were expressed only toward Jews thus the analyses performed here did not
include target and were repeated measures two way ANOVAs. For a global comparison
among the three CSE groups, in the three way ANOVAs, attention focused on the target x
responder identity interaction, and in the two way ANOVAs on the responder identity effect.
The interactions of target x responder identity x age, and responder identity x age examine the
age related differences in the three CSE groups for the two targets or one target. Significance
of between group differences was determined by Post Hock Scheffe procedure for group
comparisons, and within group patterns were examined by curve estimation regression
analyses examining linear and quadratic trends. Results that reached a significance level of .05
are reported as statistically significant.
The examination of the influence of CSE yielded consistent results. In all comparisons the
HCSE group manifested significantly higher in-group positivity and out-group negativity than
the two other groups: For positive trait attributions to the in-group, F (2,813) = 6.42, p < .05
with the means of M = 3.94, SD = .65, for the LCSE group, M = 3.99, SD = .58 for the MCSE
group, and M = 4.14, SD = .60, for the HCSE group. The difference between the LCSE and
Intergroup biases 12
the MCSE group did not reach significance. For negative trait attributions to the in-group, F
(2,464) = 47.62, p < .001 with the means of M = 2.41, SD = .74, for the LCSE group, M =
2.09, SD = .71 for the MCSE group, and M =1.74, SD = .69, for the HCSE group,. The
difference between the LCSE and the MCSE group reached significance, thus the results
single out the LCSE group as the group that attributed to the highest percent of Arabs negative
traits.
The HCSE group also attributed positive traits to significantly lowest percent of the outgroup members, F (2,816) = 73.64, p < .001, with the means of M = 2.97, SD = .68, for the
LCSE group, M = 2.59, SD = .71, for the MCSE group, and M = 2.20, SD = .74, for the
HCSE group. The difference between the LCSE group and the MCSE group reached
significance, namely, the LCSE group attributed to the highest percent of Jews positive traits.
Finally, the HCSE group attributed negative traits to significantly highest percent of the outgroup members, F (2,816) = 73.66, p < .001, with the means of M = 3.22, SD = .81, for the
LCSE group, M = 3.78, SD = .71 for the MCSE group, and M = 4.05, SD = .80, for the HCSE
group. The difference between the LCSE group and the group MCSE group reached
significance, thus again the results single out the LCSE group as the group that attributed to
the lowest percent of Jews negative traits.
The pattern of attributing positive traits to highest percent and negative traits to lowest
percent of the in-group by HCSE group and attributing positive traits to lowest percent of the
out-group and negative traits to highest percent of the out-group confirms the predictions
regarding this group. However the prediction regarding the LCSE group was not confirmed.
The LCSE group manifested lowest in-group favoritism and lowest out-group prejudice. They
attributed positive traits to lowest percent and negative traits to highest percent of in-group
members, and lowest percent of negative traits along with highest percent of positive traits to
out-group members. With one exception, (attribution of positive traits to members to the in-
Intergroup biases 13
group), the MCSE group differed significantly from the LCSE group thus holding a midway
position between the two CSE groups.
For expressed feelings and readiness for social contact, the main effects of responder
identity reached significance and yielded a similar pattern to that obtained for trait
attributions, but here the HCSE and MCSE groups did not differ and both manifested
significantly lower positivity and higher negativity toward the out-group than the LCSE
group. The findings for the different comparisons are: For positive feelings F (2,767) = 93.89,
p < .001 with means of M = 1.85, SD = 0.53, for the LCSE, M = 1.42, SD = 0.46 for the
MCSE group, and M = 1.31, SD = .0.41 for the HCSE group. For negative feelings F (2,767)
= 85.75, p < .001 with the means of M = 2.02, SD = 0.58, for the LCSE, M = 2.32, SD = 0.61
for the MCSE group, and M = 2.68, SD = 0.53, for HCSE group. For readiness for social
contact with Jews, F (2,801) = 90.00, p < .001 with the means of = 1.81, SD = 1.25 for the
LCSE, M = 0.69, SD = 0.97 for the MCSE group, and M, = 0.91, SD = 0.94, for the HCSE
group.
Positive and negative trait attributions yielded significant third order target x responder
identity x age interactions indicating significant age-related differences between and within
the examined groups: For positive attributions, F (8,780) = 11.46, p < .001, for negative
attributions F (8,784) = 10.62, p < .001. The means for these interactions are presented in
Figure 1a - d. Positive and negative feelings and readiness for social contact that were
assessed only toward Jews yielded significant second order responder identity x age
interactions. For positive feelings, F (8,767) = 6.29, p < .001, for negative feelings F (8,776) =
11.34, p < .001, for readiness for social contact F (8,801) = 9.34, p < .001. The means for
these interactions are presented in Figure 2a - c.
Insert Figure 1 & 2 about here
Intergroup biases 14
The critical age groups for significant differences among the different CSE groups are the
three younger groups, aged 8 - 9, 10 – 11, and 12 – 13, namely, the age range of 8 - 13. In
these three age groups, the differences between the high and low CSE groups reached
significance almost on all the examined variables: The 8 - 13 year-olds from the HCSE group
expressed lower positivity (Fig. 1a, Fig.2a, 2c) and higher negativity (Fig. 1c, 2b) toward the
out-group than those from the LCSE group. The 8 - 9-year-olds from the HCSE group also
expressed higher positivity toward the in-group than their counterparts from the LCSE group
(Fig. 1b). However, one exception to this pattern is the identical level of positivity toward the
in-group expressed by the 10 – 11-year olds in the two compared groups (Fig. 1b). With
regard to negativity toward the in-group (Fig. 1d) all three age groups in the HCSE group
expressed lower negativity than did those in the LCSE group. Importantly, from the three
youngest age groups, most pronounced peaks of low positivity, high negativity toward outgroup and high positivity, low negativity toward the in-group were manifested by the HCSE
12 -13-year-olds.
The MCSE group produced a less consistent pattern for the tree younger groups than the
two other groups. Here, the group aged 8 - 9, differed significantly from the HCSE group in
expressing less negative feelings toward the out-group (Fig.2b). On all other measures the 8 –
9-year-olds in the MCSE and HCSE groups were very similar. This means that at the age of 8
-9, on most measures the difference between MCSE and LCSE was also significant (Fig.1a –
d, Fig.2a, c).
From the age of 10 – 11, through 12 - 13 participants from the MCSE group tended to
abandon the extreme beliefs and attitudes characterizing the HCSE group maintaining it only
on two measures of expressed positivity toward the out-group (Fig. 2a, c), otherwise they
were closer to the LCSE group than to the HCSE group or held an in-between position that
did not differentiate them significantly from the other groups (Fig. 1a – d, Fig. 2b).
Intergroup biases 15
The groups aged 14 – 15, in the three CSE groups reached very similar level of
moderation in positivity/negativity toward the in- and out group (Fig. 1a-d, Fig 2a-c). With
two exceptions, the moderation and similarity among the CSE groups were maintained by the
16 -17-year-olds. The exceptions appeared in the MCSE group, in which the 16 - 17-year-olds
expressed significantly more negative feelings toward the out-group (Fig. 2b), and lower wish
for social contact with its members than the two other groups (Fig. 2c).
Results for age trends are presented in Table 1. As may be seen, three measures yielded
significant linear trends not accompanied by a quadratic trend (negative attributions to the outgroup in the LCSE group, positive feelings toward out-group in the MCSE, and negative
feelings in the HCSE group). In most instances either quadratic or both trends reached
significance. Since the quadratic trends account for more variance in each case and our
interest is in a particular age groups, when both trends are significant the focus is on the
quadratic trend. Inspection of the actual curves reveals different trends according to sample,
positivity/negativity and target, but almost always the peak and turning point in the curve was
at the age of 12 -13, i.e. in early adolescence.
Insert Table 1 about here
The LCSE group produced fewer significant trends than the other groups indicating more
stability over the age span than the two other groups. The significant linear trend obtained for
the LCSE group indicates that negative attributions toward the out group increased with age.
However, the highest level of their negative attributions toward the out-group reached the
lowest level of the negativity attributed by the two other groups (Fig.1c). The other significant
trends for the LCSE group include positive attributions to the in-group and negative feelings
expressed toward Jews. Both yielded quadratic trends indicating an increase from age 8 – 9 to
age 10 - 11, but whereas after this increase the positive attributions to the in-group remained
Intergroup biases 16
unchanged (Fig. 1b) the negative feelings toward the out-group continued to increase reaching
a peak at 12 -13 and only at the age of 14 -15 and 16 -17 decreased gradually (Fig. 2b).
For the MCSE group, all variables yielded significant trends most including both linear
and quadratic trends. An inspection of the curves indicates a decline in positivity toward the
in-group at age 10 -11 and maintenance of this level through the examined age trajectory (Fig.
1b). Negativity toward in-group increased with age, then, at the age of 12 – 13 decreased
forming a gentle inverted U curve, (Fig. 1d). An opposite trend emerged toward the outgroup, here a gradual increase in positivity continued till the age of 14 – 15 and decreased at
the age of 16 – 17 (Fig. 1a, Fig. 2c) reaching the lowest point at the age of 16- 17. The same
trend appeared with respect to positive feelings toward the out-group only here the trend was
linear, namely the observed reduction in this measure in the oldest age group did not change
the linear trend (2a). Negativity toward the out-group was not consistent, negative attributions
followed a gentle U pattern of a decrease and then at the age of 14 -15, an increase (Fig. 1c).
Expressed negative feelings toward the out-group formed a zigzagging pattern, first rising
with age, then at the age of 12 -13 declining and at age 14 -15 increasing and continuing to
increase reaching the highest point at the age of 16 - 17 (Fig. 2b).
The curves in the HCSE group are more distinct, and in contrast to the age variability
characterizing the MCSE group, the HCSE group displayed very consistent trends always
involving a peak in the same age group (12 - 13-year-olds). The graphs depict an increase in
positivity toward the in-group only for the 12 -13-year olds, (Fig. 1b)and a stability of a very
low negativity toward the in-group till the age of 12 -13 and then, at 14 -15, an increase (Fig.
1d).
Two measures of positivity toward the out group, (Fig. 1a, Fig. 2a,) display a stable trend
for the three younger groups and then, at the age of 14 -15 increase. The third measure of
positivity displays a more erratic pattern in the younger groups, but here more distinctly, the
Intergroup biases 17
lower point is in early adolescence and there after an increase (2c). Negative attributions to
the out- group display an inverted pattern, increasing with age reaching the highest point at
the age of 12 – 13 and then decreasing (Fig. 1c). Negative feelings represent the only
exception for this group. This measure produced a linear trend indicating a decrease with age
to a similar point reached by the LCSE group (Fig. 2b).
Discussion
The present study aimed to advance the understanding of developmental and contextual
influences, during different stages of adolescence on intergroup responses. It presents a
reexamination with Arab participants, of predictions derived from IDCT, previously
examined with Jewish Israeli participants. It was hypothesized that CSE and the experiences
related to the development of identity in pre- and early adolescence have a decisive influence
on intergroup beliefs and attitudes. Based on previous propositions and findings (Erikson,
1968; Crocker, Blaine, & Luhtanen, 1993; Crokcer & Luhtanen, 1990; Luhtanen& Crokcer,
1991; Marcia, 1980; 1998; Tajfel, 1981, Turner, 1999), we expected that irrespective of age,
and particularly in pre- and early adolescence CSE will influence intergroup attitudes. Largely
our predictions were confirmed and were in line with previous research that has examined
CSE as an individual difference measure (Crocker, Blaine, Luhtanen, 1993; Crokcer &
Luhtanen, 1990; Luhtanen& Crokcer, 1991).
Polarization and generalization
A general look at the results portrayed in Fig. 1 reveals that positive (Fig, 1a, b) and especially
negative (Fig, 1c, d) attributions toward each of the two groups are plotted on the opposite
sides of the scale. Namely, a high percent of in-group members were attributed positive traits
while only low percent were attributed negative traits. An opposite configuration is associated
with the out-group. This pattern demonstrates that conflict polarizes in- and out-group
perception homogenizing the beliefs about both groups. Further, it may be expected that the
Intergroup biases 18
negative feelings for the out group and low readiness for social contact (Fig.2b, c) will
reinforce and generalize the biases. This in turn may lead to the categorization of the outgroup as immoral, to its moral exclusion (Opotow, 1990) and delegitimization (Bar-Tal, 1989;
1990).
The same polarized characterizations of the in- and out-group, and same pattern of
attitudes were displayed by same aged Israeli Jewish participants (Bar-Tal & Teichman, 2005;
Teichman, 2001). This reciprocity demonstrates that conflicts have similar effects on biases of
the involved parties and may offer a partial explanation for the difficulties faced by groups in
conflict to resolve the controversial issues that antagonize them. Realizing that intergroup
biases are formed in early childhood may help to prevent them or to help design age-related
interventions for changing them (Bar-Tal & Teichman, 2005; Teichman & Bar-Tal, in press).
Differences between the CSE groups
HCSE participants attributed to a highest percent of in-group members positive traits, and to a
lowest percent negative traits. Toward the out-group the pattern was reversed: Significantly
lower percent were attributed positive traits and a highest percent were attributed negative
traits. The HCSE group also expressed more negative feelings and least positive feelings
toward the out-group, and lowest readiness for social contact with its members than the LCSE
group, but on these measures did not differ significantly from the MCSE group. The
consistent difference between the HCSE and LCSE groups as a whole, confirm the hypothesis
that as in minimal group situations, in a real threat situation, the HCSE group displays social
biases. The new finding in this study is that along with the pronounced in-group positivity,
HCSE participants also displayed the highest out-group negativity and lowest out-group
positivity than the LCSE and on trait attributions also than the MCSE group. Sharing a high
sense of group pride, in a real, meaningful social context (Crocker, Blaine, Luhtanen, 1993;
Intergroup biases 19
Tajfel, 1981), it appears that these participants used all the possible ways for expressing biases
and presumably for enhancing their self-worth.
The LCSE group expressed lowest in-group positivity and highest in-group negativity,
namely, as in previous studies (Crocker, Blaine, Luhtanen, 1993; Crokcer & Luhtanen, 1990;
Luhtanen& Crokcer, 1991) they did not distort in-group representation. However, since we
assessed in-group negativity separately our findings are more compelling. On the other hand,
contrary to the hypothesis, even in conflict and in a sensitive developmental stage, the LCSE
participants did not protect their esteem by displaying mark able out-group negativity or
lowest out-group positivity.
With the exception of positive attributions to the in-group, in which the MCSE group did
not differ significantly from the LCSE group in attributing to low percent of their in-group
members positive traits, in all other variations of trait attributions it differed significantly from
the two other groups. They expressed less in-group negativity, more out-group negativity and
less out-group positivity then the LCSE group, but also less positivity and more negativity to
the in-group and more positivity less negativity to the out-group than the HCSE group. With
regard to feelings and readiness for social contact the MCSE group did not differ from the
HCSE group, at the same time, both MCSE and HCSE groups differed significantly from the
LCSE group that expressed highest out-group positivity and lowest out-group negativity.
These results and the means obtained for the MCSE group for the different measures, indicate
that as hypothesized the MCSE group may be characterized as holding a midway position
with one similarity to the LCSE and three to the HCSE group. This suggests that although
attributing to less in-group members positive attributions than the HCSE group, generally,
they employ an intergroup strategy that enables a better self-enhancement than the LCSE
group, though apparently, not as effective as the HCSE group.
Differences between age groups in the compared CSE groups
Intergroup biases 20
In each of the LCSE and HCSE groups the three younger age groups resemble each other. In
the MCSE group the 8 – 9-year-olds display a different pattern, while the two following
groups (10 – 11; 12 – 13), generally resemble each other. With a few exceptions the older
groups in the three CSE groups also resemble each other. Accordingly, in comparing the
specific age groups in the different CSE groups we shall look generally at the younger and
older groups.
As in the overall comparison between the LCSE and HCSE groups when specific age
groups were compared, the differences between each of the three youngest groups were
significant. One exception diverted from this pattern in which the 10 - 11-year-olds in both
CSE groups attributed to the same percent of in- group members positive traits. This was
expected for the HCSE group less for the LCSE group. However, this one diversion from the
expected pattern does not change the general tendency obtained for the two samples in which
the younger groups in the HCSE group attributed to the in-group higher positivity and lower
negativity, and to the out-group lower positivity and higher negativity than the these age
groups in the LCSE group.
For the MCSE group, in most instances, the 8 – 9-year-olds resembled their counterparts
in the HCSE group, expressing positivity and negativity toward in- and out-group. In this age,
the similarity between the two groups exposed to conflict (HCSE and MCSE) and the
difference between them and the group with no exposure to conflict (LCSE) suggests that the
reactions of the 8 – 9-year-olds were influenced more by threat than by CSE. Indeed, in the
age of 8 - 9 identity issues are less salient, and it is possible that in the group with an
ambiguity regarding the CSE the youngest children did not internalize the satellites regarding
their collective identity and how it compares with others. Support for the interpretation that 8
– 9-year-olds react to threat rather than to esteem issues may be found in the findings that
same aged Jewish Israeli participants, who also experience high exposure to conflict, tended
Intergroup biases 21
to display similar responses toward their respective in- and out-groups to the response
obtained here for the 8 – 9-year-olds in the HCSE and MCSE groups (Teichman & Bar- Tal
2005).
Pre- and early adolescents in the MCSE group became significantly different from thosein
the HCSE group and resemble same aged groups in the LCSE group. Whereas the HCSE preand early adolescents expressed highest positivity and lowest negativity toward the in-group,
and lowest positivity and highest negativity toward the out-group, the responses of these age
groups in the MCSE group mellowed. These comparisons between pre-and early adolescents
in the three CSE groups constitutes the first support for the predicted age differences, pointing
out that pre-and early adolescents in the HCSE group differ from those in the two other
groups in the predicted way. Apparently, as hypothesized, pre-and early adolescents
belonging to HCSE groups express more biases and are able to better confront identity issues
than the other groups. The shift in the MCSE reducing their biases at pre-and early
adolescence indicates that though on some of the general comparisons they resemble the
HCSE, in the critical age groups of identity formation they resemble the LCSE group.
The results obtained for the two older groups (14 -15 and 16 -17) in the three compared
groups, generally indicate moderation on all biases. Often, the two older groups draw closer
to each other in their beliefs and attitudes. Apparently, when cognitive development advances
and secure identity begins to emerge the need for self-enhancement is less predominant and
intergroup biases tend to mellow. This consistent change that was reported also for same aged
Jewish Israeli participants (Bar-Tal & Teichman, 2005; Teichman, 2001) may be seen as an
additional confirmation for the association between insecure identity, the need for selfenhancement, and social biases.
However, the similarity between the two older age groups in the three compared groups is
not complete. In two instances, the 16 -17-year-olds in the MCSE group diverted from the
Intergroup biases 22
moderation pattern expressing more negative feelings toward the out-group and less readiness
for social contact with its members than the two other groups. This may be considered
expected responses for younger participants from a LCSE group. For the older adolescents in
the MCSE these responses may constitute a delayed attempt to protect their self-esteem.
However, since these two responses represent exceptions and changes in attitudes, unless
more direct changes toward the out-group follow (like for instance expressing negative beliefs
toward the out-group), the observed changes do not appear to suffice for protecting selfworth.
Considering the changes between the CSE groups it is evident that the consistent
difference between the younger age groups in the HCSE and LCSE groups has the main
contribution for the global differences reported above between these two groups. The unique
age patterns obtained for younger and older age groups in the MCSE group reflect that
ambiguous social identity and collective self-esteem interfere with self enhancement and
especially in the stage when it is most needed.
Age trend within the different CSE groups
The examination of the age trends in the LCSE group shows a high age stability i.e. a defused
age pattern in which the 5 age groups displayed similar responses. The significant trends were
gentle showing, a mild linear increase in the number of out-group members attributed
negative traits that in its highest point reached the point of lowest attributions made by the
other groups, a small increase in positive traits attributed to the in-group, occurring in one age
group (10 – 11), and negative feelings toward the out group mildly increasing with age, at the
age of 12 – 13 reaching a peak and then decreased. Though these trends are similar to those
manifested by the HCSE, they may be defined as too few and too mild. Thus, again it is
evident that the LCSE group does not engage in an effective self-enhancement (by expressing
in-group positivity and out group negativity) or self-protection (by expressing out-group
Intergroup biases 23
negativity) and the pre-and early adolescents in this group manifest rare and weak attempts
for such intergroup responses.
The age trends manifested by the HCSE group are consistent and the curves for this group
are distinct indicating that pre- and particularly early adolescents show the predicted trend of
high peak in in-group positivity, low peak in in-group negativity and a reversed trend for the
out-group. Occasionally, the trend in the three young groups is stable (Fig. 1a, 1d, 2a), in an
other instance there is a gradual increase and decrease (Fig1,c,) and occasionally the early
adolescents stand out (Fig. 1b, 2c). The negative feelings toward the out-group are the only
variable that produced a linear trend indicating a decrease. However, the decrease proceeds
from the maximal point possible for expressed negativity, not reached by any of the other
groups to a point still higher than the one reached by the LCSE group. It appears that in the
insecure stages of pre and early adolescence, adolescents who feel that they can draw on the
status of their group for advancing their positive identity take advantage of all available
means. In a later stage, when more secure with regard to their identity, the adolescents in this
group consistently allow themselves to express moderation in in-and out-group biases.
The age-related findings for the HCSE group replicate many of the developmental
patterns obtained for same-aged Jewish Israeli participants. The resemblance between the
results of the HCSE Palestinians and those often obtained for the Jewish participants may be
attributed to the fact that the later, being a majority group, experiencing an intractable conflict
also represents a HCSE group.
Generally, the age trends for the responses for the MCSE group reveal a pattern that may
be characterized as an opposing pattern to that manifested by the HCSE group. The
significant trends for this group indicate a decrease in positive attributions to the in-group at
the age of 10 - 11 that in the next age group is followed by an increase and then maintains
stability (Fig. 1b). This is accompanied by a mild increase in negative attributions to the in-
Intergroup biases 24
group peaking at the age of 12 - 13 and decreasing thereafter (Fig. 1d). With regard to the outgroup gradual age-related trend of increase on all measures of positivity may be observed till
the age of 14 - 15, or 16 – 17 (Fig. 1a, 2a, 2c), and a decrease in negativity peaking at the age
of 12 - 13 and increasing thereafter (1c).
Interestingly, the lowest points of in-group positivity and highest point of in-group
negativity, and lowest point of out-group negativity and highest point of out-group positivity
in the MCSE were displayed by the early adolescents and the change occurred thereafter. This
consistent trend indicates that in the MCSE group as well 12 - 13 is a critical age, but before
changing their response patterns in favor of the in-group the pre and early adolescents in this
group experiment with favoring the out-group and rejecting the in-group. These findings
demonstrate the difficulties of pre- and early adolescents in groups with LCSE or ambiguous
CSE.
The trends for the older age groups in the MCSE indicate that the significant changes
obtained in the comparison between them and same aged groups in the two other CSE groups
stem from significant changes they have undergone in relating to their group and to the outgroup, most probably attempting to cope with identity issues in a similar way to that of the
pre- and early adolescents in the HCSE group. It appears that in groups in which pre- and
early adolescents do not manifest intergroup biases, older adolescents either continue patterns
from younger ages (LCSE group), or increase biases in a later stage (MCSE group). The fact
that the older adolescents in the HCSE group reduced all types of biases suggests that
intergroup biases in pre- and early adolescence are an adaptive strategy for coping with
identity issues.
On the other hand, it may be argued that the fact that in this study the LCSE and MCSE
participants did not express biases or expressed mild and rare biases does not mean they did
not hold biases. It only indicates they could not express them directly. In line with the idea
Intergroup biases 25
offered by Aberson, Healy, and Romero (2000), it may be suggested that participants with low
or unclear collective self-esteem have similar needs to those experienced by participants with
HCSE however they need to be provided with different channels, like for instance implicit
measures, for expressing these needs. Nonetheless, overall the consistent results obtained in
this study suggest that the development of intergroup biases in adolescence is influenced by
CSE.
References
Aberson, C. L., Healy, M., & Romero, V. (2000). Ingroup bias and self-esteem: A metaAnalysis. Personality and Social psychology Review, 4, 157-173.
Aboud, F. E. (1988). Children and prejudice. Oxford, England: Basil Blackwell.
Abrams, D., & Hogg, M. A. (1988). Comments on the motivational status of self-esteem in
social identity and intergroups discrimination. European Journal of Social Psychology, 18,
317-374.
Bar-Tal, D. (1989). Delegitimization: The extreme case of stereotyping and prejudice. In D.
Bar-Tal, C. F. Graumann, A.W. Kruglanski & W. Stroebe (Eds.), Stereotyping and
prejudice: Changing conceptions (pp. 169-182). New York: Springer Verlag.
Bar-Tal, D. (1990). Causes and consequences of delegitimization: Models of conflict and
ethnocentrism. Journal of Social Issues, 46 (1), 65-81.
Bar-Tal, D. (2001). Why does fear override hope in societies engulfed by intractable conflict,
as it does in the Israeli society? Political Psychology, 22, 601-627.
Bar-Tal, D., & Teichman, Y. (2005). Stereotypes and prejudice in conflict: Representations of
Arabs in Israeli Jewish society. New York: Cambridge University Press.
Black-Gutman, D., & Hickson, F. (1996). The relationship between racial attitudes and socialcognitive development in children: An Australian study. Developmental Psychology, 32,
Intergroup biases 26
448-456.
Brewer, M. B. (1999). The psychology of prejudice: In-group love or out-group hate?
Journal of Social Issues, 55, 429-444.
Brown, R. (2000). Group processes (2nd ed). Oxford: Blackwell.
Cameron, J. A., Alvarez, J. M., Ruble, D. N. & Fuligni, A. J. (2001). Children’s lay
theories about in-groups and out-groups: Reconceptualizing research on prejudice.
Personality and Social Psychology Review, 5, 118-128.
Crocker, J., Blaine, & Luhtanen, R. (1993). Prejudice, intergroups behaviour and selfesteem: Enhancement and protection motives. In M. Hogg & D. Abrams (Eds.), Group
motivation: Social psychological perspectives (pp. 52-67). Hertfordshire, England:
Harvester Wheatsheaf.
Croker, J. & Luhtanen, R. (1990). Collective self-esteem and intergroup bias. Journal
Personality and Social Psychology, 58, 60-67
Erikson, E. H. (1968). Identity: youth and crisis. New-York: Norton.
Killen, M., Lee-Kim, J., McGlothlin, H., & Stangor, C. (2002) How children and adolescents
evaluate gender and racial exclusion. Monographs of the Society for Research in Child
Development, 67 (4), vii.
Killen, M., Margie, N.G., & Sinno, S. (2006). Morality in the context of intergroup
relationships. In M. Killen & J. Smetana (Eds.). Handbook of moral development. (pp.
155-183). Mahwah, NJ: LEA.
Kohlberg, L. (1969). Stage and sequence: the cognitive-developmental approach to
socialization. In D. A. Goslin (Ed.), Handbook of socialization theory and research.
Chicago: Rand McNally.
Intergroup biases 27
Long, K., & Spears, R. (1998). Opposing effects of personal and collective self-esteem on
interpersonal and intergroup comparisons. European Journal of social Psychology, 28, 913930.
Luhtanen, R., & Croker, J. (1990). Self-esteem and intergroup comparison: Toward a theory of
collective self-esteem. In J. Sulls & T. A. Wills (Eds.) Social comparison Contemporary
theory and research. Lawrence Erlbaum Associates, Publishers, Hilsdale, NJ. pp. 211-234.
Marcia, J. E. (1980). Identity in adolescence. In J. Adelson (Ed.), Handbook of adolescent
psychology (pp. 159-187). New York: Wiley.
Marcia, J. E. (1998). Optimal development from an Eriksonian perspective. In, Encyclopedia
of mental health (Vol. 3., pp. 29-39). New York: Academic Press.
Nesdale, A. R. (1999). Developmental changes in children’s ethnic preferences and social
cognitions. Journal of Applied Developmental Psychology, 20, 501-519.
Nesdale, D. (2000). Developmental changes in children's ethnic preferences and social
cognition. Journal of Applied developmental Psychology, 20, 501-519.
Nesdale, D., & Brown, K. (2004). Children's attitudes towards an atypical member of an
ethnic group. International Journal of Behavioral Development, 28, 328-335.
Nesdale, D., Durkin, K., Maass, A., Griffiths, J. (2004). Group status, outgroup ethnicity, and
children's ethnic attitudes. Applied Developmental Psychology, 25, 237-251.
Nesdale, D., & Flesser, D. (2001). Social identity and the development of children’s group
attitudes. Child Development, 72, 506-517.
Nesdale, D., Maass, A., Durkin, K., & Griffiths, J. (2005). Group Norms, Threat, and
Children's Racial Prejudice. Child Development. 76, 652-663.
Opotow, S. (1990). Moral exclusion and justice: An introduction. Journal of Social Issues, 46,
1-20.
Intergroup biases 28
Piaget, J. (1952). The origins of intelligence in children. New York: International Universities
Press.
Rutland, A. (1999). The development of national prejudice, in-group favouritism and selfstereotypes in British children. British Journal of Social Psychology, 38, 55-70.
Sasson, A. (2003). Self-esteem and intergroup stereotypes and attitudes. Unpublished
manuscript, Tel-Aviv University, Tel-Aviv, Israel.
Tajfel, H. (1981). Human groups and social categories. Cambridge: Cambridge University
Press.
Teichman, Y. (2001). The development of Israeli children's images of Jews and Arabs and
their expression in Human Figure Drawings. Developmental Psychology, 37, 749-761.
Teichman, Y. & Bar-Tal, D. (In press). Acquisition and development of shared psychological
repertoire intergroup in conflict. In S.M. Quintana & C. McKown (Eds.), Race, racism,
and the child. New York: Wiley.
Teichman, Y. & Zafrir, H. (2003) Images held by Jewish and Arab children in Israel of people
representing their own and the other group. Journal Cross Cultural Psychology. 34, 658676.
Turner, J. C. (1999). Some current issues in research on social identity and self-categorization
theories (pp. 6-34). In N. Ellemers, R. Spears & B. Doosje (Eds.), Social identity. Oxford,
UK: Blackwell.
Vaughan, G. M. (1987). A social psychological model of ethnic identity. In: J. S. Phinney, &
M. J. Rotheram, (Eds.), Children’s ethnic socialization. (pp. 73-91). London: Sage.
Wills, T. A. (1981). Downward comparison principles in social psychology. Psychological
Bulletin, 90, 465-498.
Intergroup biases 29
Intergroup biases 30
Figure caption:
Figure 1: Positive and negative trait attributions to Jews and Arabs (1 = 0%; 2 = 25%; 3 =
50%; 4 = 75%; 5 = 100%)
Figure 2: Positive and negative feelings expressed toward Jews and readiness for social
contact with them